Liver surgery is an increasingly common surgical procedure but still remains associated with significant morbidity and mortality.1 Postoperative pulmonary complications and intraoperative blood loss are essential determinants of postoperative outcome after liver resection.2 In this context, the recent multi-centre Intraoperative Protective Ventilation (IMPROVE) trial found that the use of a lung protective mechanical ventilation strategy [low tidal volume (6 to 8 ml kg−1 ideal body weight), moderate levels of positive end-expiratory pressure (PEEP) (6 to 8 cmH2O) and repeated recruitment manoeuvres] significantly reduced postoperative pulmonary complications compared with non-protective ventilation (higher tidal volume ventilation with zero PEEP) in patients undergoing major abdominal surgery.3 It is, however, uncertain whether high PEEP and recruitment manoeuvres should be included in a lung protective ventilation strategy for patients undergoing abdominal surgery.4 It is unfortunately also widely accepted that PEEP should not be used during liver resection on the grounds that PEEP could increase central venous and hepatic venous pressures leading to higher blood loss during hepatic transection.5–7 This is a major concern because intraoperative bleeding has consistently been associated with increased postoperative morbidity and mortality.8 There are, however, few data to support or refute the hypothesis of a causal relationship between the use of PEEP and intraoperative bleeding during liver surgery. Given this uncertainty, we performed a post-hoc analysis of the IMPROVE trial with the aim of assessing whether a lung protective mechanical ventilation strategy was associated with an increase in intraoperative bleeding in the subgroup of patients that underwent liver surgery.
All patients from the IMPROVE trial who underwent liver surgery were analysed in our post-hoc study. In the multi-centre, double-blinded, stratified, parallel group IMPROVE trial, patients with an intermediate-to-high risk of postoperative pulmonary complications scheduled to undergo laparoscopic or non-laparoscopic elective major abdominal surgery with an expected duration of at least 2 h were randomly assigned to receive lung protective mechanical ventilation (tidal volume of 6 to 8 ml kg−1 predicted body weight, PEEP of 6 to 8 cmH2O and repeated recruitment manoeuvres of 30 cmH2O every 30 min) or non-protective mechanical ventilation (tidal volume of 10 to 12 ml kg−1 predicted body weight with zero PEEP and no recruitment manoeuvres).3 Patients were ineligible if they had a BMI of 35 or less or had a history of respiratory failure, sepsis or mechanical ventilation within the 2 weeks preceding surgery. Decisions about all other aspects of patient care during the intra- and postoperative periods, including anaesthetic management (type of anaesthesia, drug choices, haemodynamic monitoring, fluid management and transfusion), were left at the discretion of the attending physician according to the expertise of the staff at each centre and routine clinical practice. The primary outcome was a composite of major pulmonary and extrapulmonary complications occurring within the first 7 days following surgery. The trial was approved for all participating centres by a central ethics committee (Comité de protection des Personnes Sud-Est I, Saint Etienne, France, Chairperson F.Faisan) on 30 December 2010 according to French legislation. Written informed consent was obtained prior to randomisation from each participant.
In the present study, we assessed differences in the intervention effect (lung protective vs. non-protective ventilation) in terms of intraoperative bleeding. The primary endpoint was the total volume of intraoperative blood loss, which was recorded systematically and simultaneously by the anaesthetic and the surgical teams at the end of the surgical procedure by taking into account surgical sponges, suction canisters and the cell salvage device (if used). Secondary endpoints included the number of patients transfused, the median number of blood red cell units transfused, postoperative morbidity according to the Clavien–Dindo classification,9 length of stay in the ICU, length of hospital stay and in-hospital mortality. All the data used for this study had been collected prospectively and were made available from the original database. When required, additional data on liver function were obtained in each participating centre from patients’ medical and surgical records including: preoperative and postoperative liver tests, detailed information on surgical procedure (type of surgical intervention, number of resected segments, type of vascular clamping and duration) and anaesthetic management (anaesthetic agents, use of central venous catheter, central venous pressure and/or cardiac output monitoring, recruitment manoeuvres during hepatic transection, use of vasopressors and fluid infusion). All additional data were collected by research staff blinded to the attribution group.
Statistical analysis was performed using Stata 13 software (StataCorp LP, College Station, Texas, USA). The tests were two sided, with a type I error set at α = 0.05. Patient's characteristics were presented as mean (±SD) or median (interquartile range) for continuous data according to statistical distribution (assumption of normality assessed using the Shapiro–Wilk test) and as number of patients (proportion) for categorical variables. Comparisons between the independent groups were made using χ2 squared or Fisher's exact tests for categorical variables, and Student's t-test or the Mann–Whitney test for quantitative variables (homoscedasticity verified using Fisher–Snedecor test). Owing to this statistical distribution, the primary outcome was assessed with a non-parametric Mann–Whitney test. Considering this work as an ancillary analysis of the IMPROVE study, no formal sample size estimation was performed.
A total of 96 patients (24%) from the original IMPROVE trial underwent hepatic surgery. Seventeen patients were secondarily excluded as they had not exclusively undergone hepatic resection (n = 15) or because of missing data (n = 2). Thus 79 patients were considered for the final analysis with 41 of them randomised to the non-protective ventilation and 38 to the lung protective ventilation group (Fig. 1). Baseline characteristics were similar between the two groups (Table 1). Patients were predominately male with a mean age of 61 years. Overall 11% of patients had a diagnosis of cirrhosis and 66% underwent surgery for a carcinoma. There were no statistically significant differences regarding preoperative coagulation tests or liver tests between the groups (Table 2) although baseline haemoglobin level was statistically greater in the lung protective ventilation group. Surgical procedures are detailed in Table 3. Surgery was achieved via laparotomy in more than 90% of the cases and 37% of all patients underwent a major hepatic resection (defined as resection of ≥ three segments). There were no statistically significant differences between groups in the number of resected segments, incidence or type or duration of vascular clamping and duration of surgery. The anaesthetic management was similar in both groups regarding the use of a central venous catheter, measurement of central venous pressure and the choice of inhalational anaesthetic agent. Of note, intraoperative measurement of central venous pressure was rarely used in both groups (8.9%). Mean plateau pressures were similar in the two groups at the beginning of surgery (16.4 ± 3.2 vs. 15.2 ± 2.0 cmH2O, P = 0.08 in non-protective and lung protective groups, respectively) but were significantly higher in the non-protective group at the end of surgery (17.0 ± 3.0 vs. 15.6 ± 2.2 cmH2O, P = 0.03). The amount of intraoperative bleeding was 500 (200 to 800) ml in the non-protective ventilation group and 275 (125 to 800) ml in the lung protective ventilation group (P = 0.47). The distribution of bleeding volume is shown in Fig. 2. Fourteen (35.0%) patients received a blood transfusion in the non-protective ventilation group compared with eight patients (21.5%) in the lung protective ventilation group (P = 0.17). The number of packed red blood cells units transfused was 2.5 (2 to 4) in the non-protective ventilation and 3 (2 to 6) in the lung protective ventilation group (P = 0.54). Regardless of the allocation group, the amount of bleeding was significantly greater in major compared with minor liver resections (1060 ml ± 810 vs. 407 ± 500, P < 0.01) Haemoglobin level, coagulation tests and serum liver enzymes were not statistically different in the immediate postoperative period in the two study groups (Table 4). At 48 h postoperatively, the bilirubin level was greater (49 ± 69 vs. 21 ± 15 μmol l−1, P = 0.03) and the serum bicarbonate level lower (24 ± 4 vs. 26 ± 3 μmol l−1, P = 0.02) in the non-protective ventilation group. The postoperative haemoglobin level was similar in both groups. There was no statistically significant between groups in terms of postoperative complications according to Clavien–Dindo classification (Table 5). There was no statistically significant difference in the intervention effect in term of major pulmonary and extrapulmonary complications within the first 7 days after surgery (the primary outcome measure of the IMPROVE trial), which occurred in 24.4 and 13.2% of patients assigned to the non-protective and lung protective ventilation groups, respectively (P = 0.20). Moreover, there were no significant differences in the ICU length of stay (1.6 ± 2.6 vs. 1.7 ± 2.7 days, P = 0.90), length of hospital stay (14 ± 9 vs. 12 ± 5 days, P = 0.59) and in-hospital mortality (2.5 vs. 7.9%, P = 0.35) in the non-protective and the lung protective ventilation groups, respectively.
In this post-hoc analysis of data from the multi-centre IMPROVE trial, the use of lung protective ventilation combining low tidal volumes, PEEP between 6 and 8 cmH2O and repeated recruitment manoeuvres was not associated with an increased amount of intraoperative bleeding in patients undergoing liver resection.
After becoming a standard of care in critically ill patients, there is a growing body of evidence that lung protective mechanical ventilation strategies might be associated with better clinical outcomes in patients scheduled for high-risk surgery.10 The IMPROVE study has recently provided evidence that a strategy combining low tidal volume, PEEP and recruitment manoeuvres during high-risk abdominal surgery, compared with non-protective ventilation (higher tidal volume, no PEEP and no recruitment manoeuvres) resulted in fewer postoperative complications and reduced healthcare utilisation.3
However, in the specific field of hepatic surgery, the benefits of lung protective ventilation with PEEP during hepatic surgery might be open to question. Indeed, there are some concerns about the fact that PEEP may increase intraoperative bleeding.11 Consequently, ventilator settings including PEEP and recruitment manoeuvres were classically not recommended during liver surgery since they may have been associated with an increase in pulmonary thoracic pressure that might subsequently increase central venous pressure and intraoperative bleeding.12 However, the rationale for this is theory is relatively limited. Direct transmission of alveolar pressure to blood central venous pressure is poorly described in non-severely ill patients.13 Moreover, existence of sphincters at the suprahepatic and inferior vena cava junction14 may preclude the pressure transmission from the central venous compartment to the hepatic capillaries.15 A recent study by Lasndorp et al.16 has observed that, besides the effect of PEEP, an increase in tidal volume was also associated with an increase in airway pressure and in central venous pressure. These variations might also account for the risk of increased bleeding during hepatic transection. Of note, the mean plateau pressure was higher in the high tidal volume group at the end of the surgery. Unfortunately, monitoring of central venous pressure has been performed in a too small percentage of patients to use these data. However, these physiological considerations are consistent with our results suggesting that the application of moderate PEEP levels of 6 to 8 cmH2O, in combination with periodic recruitment manoeuvres and low tidal volume, was not associated with increased intraoperative bleeding when compared with high tidal volume and no PEEP. It is not possible to distinguish the specific effects of each element in the protective ventilation bundle in terms of blood loss as they have not been compared with one another.
The population of patients undergoing liver surgery in our study had very common surgical characteristics of modern hepatic surgery when considering the type of surgery and the number of major resections.1 Surgical techniques associated with a reduction in intraoperative bleeding such as vascular clamping17 and use of transection devices18 were also commonly used. The volume of intraoperative bleeding in the current study was within the usual range quoted in the literature, as was the 25% rate of blood transfusion. Moreover, the overall incidence and the distribution of postoperative morbidity was in accordance with the previously reported results.19 This combination of factors suggests that our results have a good external validity. Additionally and very specific to liver surgery, it might be interesting to consider that serum levels of bilirubin were significantly lower in the lung protective ventilation group at 48 h postoperatively. This result could be a positive indication of a better recovery in hepatic function in the lung protective group,20 possibly linked with reduced systemic inflammation, but this needs to be confirmed prospectively.
The results of this study come with some inevitable limitations. First, the analysis was planned post-hoc, so the results cannot be considered confirmative. Second, one might question whether the use of blood transfusion, instead of our primary endpoint (the volume of intraoperative blood loss), would have been more relevant because accurate estimation of the bleeding volume may be somewhat difficult.21 However, the volume of blood losses was reported both by the anaesthetic and surgical teams suggesting a good degree of reliability. Furthermore, blood transfusion is becoming a relatively scarce procedure during elective liver surgery with a great interinstitutional variability.22
Third, randomisation was not stratified to the type of surgery. Nevertheless, there was no imbalance in baseline patient characteristics in both groups, at least for factors related to the amount of bleeding in the surgical procedures11,23 or the anaesthetic management.24 The number of resected segments, vascular clamping, surgical duration, use of central venous catheters and central venous pressure monitoring were similar in both groups. However, it was not possible to determine the impact of ventilator strategies in more specific types of hepatectomy such as segment I resection.
Fourth, our sample size was relatively small suggesting that complementary data are needed to confirm our results, particularly concerning the blood transfusion rates. However, as indicated by Feise, it is important not to focus on statistical significance alone but also upon the quality of research within the study and the magnitude of improvement.25 The strengths of this study are the strict methodology and quality of the data of the original trial, which minimise the risk of bias. The lack of statistical power to detect an intervention effect in term of intraoperative bleeding can be challenged for two main reasons. First, the distribution of bleeding, as shown in Fig. 2, was very similar in each group with a non-statistically significant trend toward lower bleeding in the lung protective group. This difference of 225 ml, with overlapping interquartile ranges, was not numerically large and did not seem clinically relevant. Second, ex post facto calculation, when considering the non-parametric distribution of bleeding, showed that a sample size of at least 700 patients per group would have been necessary to detect a non-clinically relevant difference of 200 ml blood loss for a type I error α = 0.05 and a statistical power of 90%.26 A sample size of this magnitude for such a small clinical difference seems completely inconsistent.
In conclusion, recognising the methodological limitations of the study, the use of moderate levels of PEEP, repeated recruitment manoeuvres and low tidal volume during lung protective ventilation was not associated with increased intraoperative bleeding compared with non-protective high tidal volume ventilation during hepatic surgery.
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship: none.
Conflicts of interest: EF reports receiving consulting fees from General Electric Medical Systems, lecture fees from Fresenius Kabi and reimbursement of travel expenses from Fisher and Paykel Healthcare. SJ reports receiving consulting fees from Dräger France and Maquet France, lecture fees from Fisher and Paykel Healthcare, Abbott, and Philips and reimbursement of travel expenses from Pfizer. CP-B reports receiving consulting fees from Fresenius Kabi, Medicines company, lecture fees and reimbursement of travel expenses from Astellas, and payment for the development of educational presentations from LFB Biomedicaments, Astellas, Merck Sharp & Dohme.
Presentation: preliminary data for this study were presented as an oral communication at the SFAR congress (Société Française d’Anesthésie Réanimation), 18–20 September 2014, Paris.
1. Farges O, Goutte N, Bendersky N, et al. ACHBT-French Hepatectomy Study Group. Incidence and risks of liver resection: an all-inclusive French nationwide study. Ann Surg
2. De Boer MT, Molenaar IQ, Porte RJ. Impact of blood loss on outcome after liver resection. Dig Surg
3. Futier E, Constantin J-M, Paugam-Burtz C, et al. A trial of intraoperative low-tidal-volume ventilation in abdominal surgery. N Engl J Med
4. Hemmes SNT, Gama de Abreu M, Pelosi P, et al. PROVE Network Investigators for the Clinical Trial Network of the European Society of Anaesthesiology. High versus low positive end-expiratory pressure during general anaesthesia for open abdominal surgery (PROVHILO trial): a multicentre randomised controlled trial. Lancet
5. Johnson M, Mannar R, Wu AV. Correlation between blood loss and inferior vena caval pressure during liver resection. Br J Surg
6. McNally SJ, Revie EJ, Massie LJ, et al. Factors in perioperative care that determine blood loss in liver surgery. HPB
7. Li Z, Sun Y-M, Wu F-X, et al. Controlled low central venous pressure reduces blood loss and transfusion requirements in hepatectomy. World J Gastroenterol WJG
8. Ercolani G. Predictive indices of morbidity and mortality after liver resection. Ann Surg
9. Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg
10. Serpa Neto A, Cardoso SO, Manetta JA, et al. Association between use of lung-protective ventilation with lower tidal volumes and clinical outcomes among patients without acute respiratory distress syndrome: a meta-analysis. JAMA J Am Med Assoc
11. Sima CS, Jarnagin WR, Fong Y, et al. Predicting the risk of perioperative transfusion for patients undergoing elective hepatectomy. Ann Surg
12. Melendez JA, Arslan V, Fischer ME, et al. Perioperative outcomes of major hepatic resections under low central venous pressure anesthesia: blood loss, blood transfusion, and the risk of postoperative renal dysfunction. J Am Coll Surg
13. Yang Z, Zhou J, Sun B, et al. [The influence of positive end-expiratory pressure on central venous pressure in patients with severe craniocerebral injury]. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue
14. Lautt WW, Legare DJ, Greenway CV. Effect of hepatic venous sphincter contraction on transmission of central venous pressure to lobar and portal pressure. Can J Physiol Pharmacol
15. Brienza N, Ayuse T, O’Donnell CP, et al. Regional control of venous return: liver blood flow. Am J Respir Crit Care Med
16. Lansdorp B, Hofhuizen C, van Lavieren M, et al. Mechanical ventilation-induced intrathoracic pressure distribution and heart-lung interactions*. Crit Care Med
17. Richardson AJ, Laurence JM, Lam VWT. Portal triad clamping versus other methods of vascular control in liver resection: a systematic review and meta-analysis. HPB
18. Richter S, Kollmar O, Schuld J, et al. Randomized clinical trial of efficacy and costs of three dissection devices in liver resection. Br J Surg
19. Breitenstein S, DeOliveira ML, Raptis DA, et al. Novel and simple preoperative score predicting complications after liver resection in noncirrhotic patients. Ann Surg
20. Reissfelder C, Rahbari NN, Koch M, et al. Postoperative course and clinical significance of biochemical blood tests following hepatic resection. Br J Surg
21. Toledo P, McCarthy RJ, Hewlett BJ, et al. The accuracy of blood loss estimation after simulated vaginal delivery. Anesth Analg
22. Ozier Y, Pessione F, Samain E, et al. French Study Group on Blood Transfusion in Liver Transplantation. Institutional variability in transfusion practice for liver transplantation. Anesth Analg
23. Clavien P-A, Petrowsky H, DeOliveira ML, et al. Strategies for safer liver surgery and partial liver transplantation. N Engl J Med
24. Scalzone R, Lopez-Ben S, Figueras J. How to transect the liver? A history lasting more than a century. Dig Surg
25. Feise RJ. Do multiple outcome measures require p-value adjustment? BMC Med Res Methodol
26. Noether GE. Sample size determination for some common nonparametric tests. J Am Stat Assoc